Original Article
Altruism toward in-group members as a reputation mechanism

https://doi.org/10.1016/j.evolhumbehav.2009.09.004Get rights and content

Abstract

To test the hypothesis that sensitivity to monitoring drives people to act altruistically toward members of their own community, two experiments investigated whether an eye-like painting promotes altruism toward in-group members, but not toward out-group members. Participants played the role of dictator in a dictator game with another participant (a recipient) who was from the minimal in-group or out-group. Participants knew whether their recipient was an in-group member or an out-group member, but were informed that their recipient did not know the group membership of the dictator. In-group favoritism occurred only when participants were facing a computer desktop which displayed a painting of eyes, but did not occur in the absence of eyes. These findings demonstrate that the eye painting displayed on the participant's computer screen worked as a cue for monitoring and thus enhanced the participant's altruistic behavior.

Introduction

Altruism toward genetically unrelated individuals whom one is not likely to meet again is an evolutionary puzzle. This is because the standard explanations of altruism among animal species, including kin selection (Hamilton, 1964) and reciprocal altruism (Trivers, 1971), are incapable of explaining altruism toward genetically unrelated individuals whom one is not likely to interact with again. From an evolutionary perspective, such altruistic behavior seems to be, at least on the surface, only detrimental to one's inclusive fitness. And yet, such behavior is not unusual; many people anonymously donate money to various charities, and participants in both laboratory and field experiments who are strangers to each other behave in altruistic and cooperative manners even in one-shot encounters (Berg et al., 1995, Fehr and Fischbacher, 2003, Forsythe et al., 1994, Gintis et al., 2003, Henrich et al., 2006, Hoffman et al., 1996, Ledyard, 1995, Messick and McClintock, 1968). In this study, we present experimental evidence in support of the proposal that such altruism can be understood as an implicit strategy to adapt to a system of indirect reciprocity.

The growing literature on indirect reciprocity in theoretical biology aims to provide an answer to this evolutionary puzzle. Altruistic behavior toward someone whom one never sees again will not be rewarded by the recipient of such behavior. This effectively eliminates the possibility for altruists to acquire gains through mutually altruistic behavior, that is, what Trivers (1971) calls reciprocal altruism. This, however, does not eliminate the possibility that altruists are indirectly reciprocated by other altruists. When other altruists are conditional altruists, or those who behave altruistically only toward other altruists but not toward egoists, it is possible for altruists to acquire more benefits than egoists. Thus, altruism can evolve when there are a sufficiently large number of conditional altruists in a community, such that the extra benefits an altruist receives from conditional altruists exceed the cost of behaving altruistically toward others.

It should be noted, however, that conditional altruism can evolve only under specific conditions. Several models have been proposed as candidates for specifying the exact nature of these conditions (e.g., Leimar and Hammerstein, 2001, Nowak and Sigmund, 1998, Panchanathan and Boyd, 2004, Takahashi and Mashima, 2006). In this article, we do not discuss the similarities and differences among those models. Instead, we would like to point out that, despite some differences in the specification of the conditions for being a recipient of conditional altruism, information about one's past behavior toward others plays a crucial role in the success of all indirect reciprocity models. Unless one's altruistic or egoistic behavior toward others is known to the other conditional altruists, conditional altruism cannot function. In other words, reputation about one's behavior plays an indispensable role for the success of conditional altruism as a fitness-enhancing strategy.

In all of the models of indirect reciprocity cited above, transparency of information about all community members' past behavior is assumed, although the models differ from each other in the nature of the information available to other members. For example, the “image scoring” model proposed by Nowak and Sigmund (1998) requires the availability of information about others' immediate past — i.e., whether each of the other members of the community acted altruistically or not in the immediately preceding trial. The other models, such as the “standing” model proposed by Leimar and Hammerstein (2001), the “extra standing” model and the “strict discriminator” model proposed by Takahashi and Mashima (2006), require further information; the players in these models are assumed to know, for example, whether someone behaved altruistically toward someone else in the previous trial, and, furthermore, whether that someone else behaved altruistically toward yet another person two trials ago. Despite these differences, a fundamental assumption shared by all of these models is that all players' immediate past behavior is known to all of the other players. In this study, we seek to examine the psychological implications of this fundamental assumption used in various models of indirect reciprocity.

One immediate psychological implication of the above assumption is that humans are sensitive to how others behave toward members of their own community. In their cheater detection studies, Cosmides and Tooby (1989) have demonstrated that humans are sensitive to others' behavior, especially if others' behavior does not meet the moral standard of the community. Wang (1996) further demonstrated that such sensitivity to other members' behavior is stronger in smaller groups than in larger groups.

What we want to pursue in this study is another psychological implication of the above assumption, which is as obvious as the first one, but has not been fully explored yet. That is, humans are sensitive to the fact that their behavior is being observed by other community members. People must behave in a way to enhance their reputation of behaving altruistically toward their community members in order to qualify as someone who deserves similar treatment from other members of the community. The fact that people generally behave in a more cooperative and altruistic manner when their behavior is observed as opposed to when under complete anonymity has been repeatedly demonstrated in experimental studies. For example, face-to-face contact between game players (a situation where players can observe each other's behavior) has been repeatedly demonstrated to enhance cooperation level in prisoner's dilemma and other related games (e.g., Bixenstine et al., 1966, Brechner, 1977, Dawes et al., 1977, Kurzban, 2001). In other studies, participants were found to behave in a more altruistic manner when they expected that their reputation would spread to other participants (Barclay, 2004, Barclay and Willer, 2007, Milinski et al., 2002).

Accumulating evidence suggests that people are sensitive not only to the explicit information that their future rewards depend on how others perceive and judge their behavior, but also to subtle cues suggesting that they may be monitored by others. A good example is found in a study by Bateson, Nettle and Roberts (2006), which demonstrated that users of a common coffee machine in an office voluntarily donated more money to maintain the supply of coffee when a poster of eyes was posted on the wall than when a poster of flowers was posted. The presence of the eyes is assumed to function as a cue to the presence of monitoring in this situation, thus enhancing the altruistic behavior. Burnham (2003) demonstrated that presenting a picture of the recipient to the dictator in a dictator game prompted the dictator to allocate more money to the recipient. In another experiment of a dictator game, Haley and Fessler (2005) found that participants who played the role of a dictator voluntarily gave more share of an endowment of $10 when their computer screen displayed a painting of eyes. Burnham and Hare (2007) added further evidence to the effect of eyes on a computer screen by presenting a robot figure which appeared to be looking forward. They found that the presence of the robot on the computer screen promoted the cooperation level in a public goods game. Findings in these studies suggest that the presence of eye figures functions as a cue to the operation of monitoring in the situation, prompting participants to act in a more altruistic manner. Furthermore, Rigdon, Ishii, Watabe and Kitayama (2009) found that even an abstract figure remotely resembling a human face can have the same effect in a dictator game experiment in which participants were presented with three black dots in either the shape of a triangle, “∴”, or an upside-down triangle. In their study, male participants allocated more money to the recipient when they saw an upside-down triangle of three dots, which remotely resembles human face, than when they saw a straight triangle. These findings provide support to the view that at least some of prosocial behavior frequently observed in one-shot games is a product of intuitive decision making driven by subtle but salient cues. Such a cue-driven psychological mechanism operates independently of the process of explicit decision making in which relevant pieces of information are deliberately evaluated and compared.

As shown above, we already have substantial evidence that humans are sensitive to cues suggesting the presence of monitoring and respond to such cues in the form of enhanced altruistic behavior. The goal of the current study was to add a new piece of evidence showing that such sensitivity to reputation operates mostly within one's own group where one's reputation matters most. Such sensitivity to monitoring by others should be heightened when those observing one are members of one's own community, rather than strangers. This is because the information about one's behavior held by a stranger is not very likely to be transmitted to someone with whom one may later interact. In contrast, if the same piece of information is held by a member of one's own community, it is more likely to be transmitted to other community members with whom one is likely to interact with.

In explaining in-group favoring reward allocation in a dictator game, Yamagishi and Mifune (2008) argued that altruistic behavior is fitness enhancing when it is directed toward, and is observed by, community members, as only the community members are in a position to repay altruistic acts through indirect reciprocity. In other words, behaving altruistically toward outsiders to their own community or being observed by outsiders does not qualify the altruist as a recipient of altruistic behavior by other community members. In order to test this argument, they compared two conditions concerning the commonality of group membership information. In their study, participants were first divided into one of two minimal groups (Tajfel, Billig, Bundy & Flament, 1971) and played a dictator game as a dictator. Minimal groups are artificially created in the laboratory by categorizing participants into one of several groups based on a trivial criterion such as preference for painters. No communication or interaction within and across groups is allowed. In one condition, the recipient of the money the dictator allocated was an in-group member. In the other condition, the recipient was an out-group member. These two conditions were crossed by another experimental manipulation concerning the knowledge the two players had about their group membership. In the “common knowledge” condition, both the participant (dictator) and the recipient mutually knew which group they both belonged to. In the private knowledge condition, only the participant knew which group they both belonged to and the recipient did not know which group the dictator (i.e., the participant) belonged to. Yamagishi and Mifune predicted that participants would allocate more money to an in-group recipient than to an out-group recipient only in the common knowledge condition. According to the logic presented earlier, being identified as a community member by other community members is a necessary condition for the reputational function of acting altruistically to operate. This was the exact situation in which Yamagishi and Mifune's participants faced in the common knowledge condition. This explains why participants behaved more altruistically toward in-group members than toward out-group members. On the other hand, since participants in the private knowledge condition were not identified by their own group member, making the situation irrelevant to reputation concerns, no altruism enhancement toward in-group members was found in this condition. Similar results have been repeatedly demonstrated in simultaneously played prisoner's dilemma games (Jin and Yamagishi, 1997, Kiyonari, 2002, Yamagishi et al., 1999, Yamagishi et al., 2005, Yamagishi et al., 2008), a sequentially played prisoner's dilemma game (Horita & Yamagishi, 2007) and some variations of the trust game (Foddy et al., 2009, Kiyonari et al., 2007, Suzuki et al., 2007).

The purpose of this study was to provide further evidence to support the argument of Yamagishi and Mifune (2008) that in-group favoring behavior that was observed in the minimal group settings can be best explained as an adaptive response to the indirect reciprocity environment. In this article, we used the private knowledge condition that has been reported to produce no in-group favoring effect as a control and manipulated cues of monitoring (presence vs. absence of stylized-eye image). We seek to demonstrate that displaying the stylized eyes on the participant's computer screen would have a similar altruism-enhancing effect to commonality of knowledge in Yamagishi and Mifune's (2008) study.

If in-group favoring behavior in the minimal group setting is driven by a psychological mechanism that makes people adapt to indirect reciprocity environments, participant's sensitivity to monitoring should be stronger when they deal with community members (in-group) than when they deal with someone who is known to be a stranger (out-group). According to this logic, we expect that, even in the private knowledge condition, the altruism-enhancing effect of eye figures demonstrated by Haley and Fessler (2005) should be more likely to occur when participants play a dictator game with an in-group recipient than when they play the game with an out-group recipient.

Section snippets

Study 1

Two experiments were conducted. The goal of the first experiment was to demonstrate that in-group favoring behavior occurs even in the private knowledge condition when participants face a salient cue of monitoring. Both experiments used only the private knowledge condition in which the previous study (Yamagishi & Mifune, 2008) found no in-group favoring behavior; the common knowledge condition was not used in either of the two experiments. We did not include in the first study the condition in

Study 2

The design of the second experiment was the same as that of the first experiment, except for the following three aspects. First, we added a second condition of no eyes present. Second, whereas the allocators' choice was limited to three options in the first study, in the second study, the allocators were allowed to divide the money any way they preferred (in increments of 10 yen). Third, instead of using only kabuki eyes as in the first study, we also used the Horus-eye figure adopted from

General discussion

The hypothesis concerning the effect of eyes, supported by the data from this study, is based on the fundamental assumption that sensitivity to monitoring by community members is a proximate mechanism to promote human adaptation to group life. We argued in the introduction that the proximate mechanism of human sensitivity to monitoring is based on the social mechanism to promote mutual cooperation within the group. We further argued that this social mechanism is based on the principle of

Acknowledgments

We would like to thank colleagues at Hokkaido University who generously let us recruit potential participants from their classes. We also thank Victoria Yeung and Joanna Schug for helping us improving the manuscript.

References (50)

  • KoopmansR. et al.

    Collective action in culturally similar and dissimilar groups: An experiment on parochialism, conditional cooperation, and their linkage

    Evolution and Human Behavior

    (2009)
  • MessickD.M. et al.

    Motivational bases of choice in experimental games

    Journal of Experimental Social Psychology

    (1968)
  • RigdonM. et al.

    Minimal social cues in the dictator game

    Journal of Economic Psychology

    (2009)
  • ShinadaM. et al.

    False friends are worse than bitter enemy: ‘altruistic’ punishment of in-group members

    Evolution and Human Behavior

    (2004)
  • TakahashiN. et al.

    The importance of subjectivity in perceptual errors on the emergence of indirect reciprocity

    Journal of Theoretical Biology

    (2006)
  • WangX.T.

    Domain-specific rationality in human choices: Violation of utility axioms and social contexts

    Cognition

    (1996)
  • BarclayP. et al.

    Partner choice creates competitive altruism in humans

    Proceedings of the Royal Society B - Biological Sciences

    (2007)
  • BatesonM. et al.

    Cues of being watched enhance cooperation in a real-world setting

    Biology Letters

    (2006)
  • BixenstineV.E. et al.

    Collaboration among six persons in a prisoner's dilemma game

    Journal of Conflict Resolution

    (1966)
  • BurnhamT.C. et al.

    Engineering human cooperation — Does involuntary neural activation increase public goods contributions?

    Human Nature

    (2007)
  • DawesR. et al.

    Behavior, communication and assumptions about other peoples' behavior in a commons dilemma situation

    Journal of Personality and Social Psychology

    (1977)
  • FehrE. et al.

    The nature of human altruism

    Nature

    (2003)
  • FehrE. et al.

    Altruistic punishment in humans

    Nature

    (2002)
  • FoddyM. et al.

    Group-based trust in strangers: the role of stereotypes and expectations

    Psychological Science

    (2009)
  • FromingW.J. et al.

    Prosocial self-schemas, self-awareness, and children's prosocial behavior

    Journal of Personality and Social Psychology

    (1998)
  • Cited by (0)

    The studies reported in this article were supported by Grant-in-aid 19330139 and 08J05455 by the Japan Society for the Promotion of Science.

    View full text